Python 在没有keras的情况下，如何计算tensorflow上的列车和验证损失？

作为第一步，我想在每个历元后使用tensorflow复制keras损失和验证损失打印值。我当前的代码如下所示：

optimizer = tf.train.AdamOptimizer(0.01)

X = tf.placeholder("float", [None, num_features])

y_pred = autoencoder_model #This is my model with layers, weights and biases

y_true = X

loss = tf.reduce_mean(tf.pow(y_true - y_pred,2))

init = tf.global_variables_initializer()

with tf.Session() as sess:
sess.run(init)

for epoch in range(training_epochs):
    for batch in range(bathces):
        batch_x = data[batch]

        _, c = sess.run([optimizer, loss], feed_dict = {X: batch_x})

    print('loss is {}'.format(c))
    #### print('val loss is {}'.format()) Need to print validation loss here